Previous - Next
Part II: Measurement of rainfall-induced erosion processes
1. General context
2. Measuring design
3. Measurement programme
4. Data processing and presentation
1.1 Hitherto experience in the measurement of rainfall-induced erosion processes in the region
1.2 Justification for implementing measurements of erosion processes
1.1 Hitherto experience in the measurement of rainfall-induced erosion processes in the region
The complex nature of the erosion phenomena conditions from the first stage any measurement activity. Which part of the erosion process is going to be measured? Is the interest focused on the on-site process of soil particle entrainment and incipient transport, or are we interested in the sediment delivery of the detached materials through the drainage network? Do we want to monitor channel erosion, characterized by thalweg and bank scour deposition cycles, or, are we even interested in the complete erosion cycle?
A particular, specific, experimental response corresponds to each of these questions. Such response is formulated in terms of the research project, which is characterized by a particular field experimental design and data acquisition programme. It is essential to clarify from the beginning which part of the erosion cycle is the subject of our interest and why.
Naturally, it is the research objective arising from the particular management question that has to be solved, that determines that part of the erosion cycle that will be measured.
Consequently, there are no established rules to guide the design of the experimental facilities and data acquisition programme, which additionally are strongly conditioned by the natural features of the study area.
Nevertheless the experience and the rationale of erosion measurements allows two categories of erosion phenomena measurement to be identified:
a) On site erosion measurement, and
b) System level measurement.
On site measurement monitors the erosion phenomena at the level of soil processes. It can thus focus on the measurement of splash, sheet and rill erosion.
The typical experimental framework is the erosion plot. Instrumentation consists of a rainfall measurement device, pluviometer or pluviograph, and a sediment measurement device, that collects sediment produced in the plot and generally stores a known portion of it to be measured and eventually analyzed. The most common devices are the multi-slot divisor and the Coshocton wheel.
Erosion nails and micro profile sticks are also used to establish fixed reference points on the soil surface, so that elevation changes can be related to the amount of material removed and eventually deposited on other sectors of the plot, before being collected at the outlet.
System measurement refers to the monitoring of the erosion cycle at the watershed level. In this category are several measurement schemes with different degrees of detail and comprehensiveness of the monitoring of the erosion cycle that is taking place in the selected watershed. From a minimum scheme, that simply records the suspended sediment delivered at the watershed outlet, to a complete experimental design which implies the monitoring of the on-site processes, sometimes in several plot groups inside the watershed, and the measurement of the suspended (wash and suspended load) and scoured (bed load) sediment, there are several possibilities.
The instrumentation used for system level measurement are diverse and are described in detail in Chapter 3.
Naturally no qualification or assessment should be made in reference to the scheme. The more complex and precise schemes are not necessarily the best. The quality of the experimental scheme is strictly related to the research objectives being met and to the capacity to clearly answer the management related questions. The experimental response must fit this requirement. Some questions can be answered through plots (i.e. the sheet and rill erosion occurring in relation to particular types of land-use and site conditions) and some others require a complex network of experimental watersheds to reach only an approximate answer (i.e. the modeling of the sediment delivery, that has to be made in order to make predictions of the amount of eroded sediments reaching a coastal sector).
There are several experimental areas where erosion measurements are currently being made in the Mediterranean. In addition to those specifically related to this project, Vallcebre (Spain), Esen (Turkey) and Oued Ermel (Tunisia), which are comprehensively described in (PAP/RAC-UNEP, 1997), others are currently in operation and up to date in terms of their scientific objectives and instrumentation. Amongst these can be mentioned the experiments at Rambla Honda, Almeria (Spain), managed by the EEZA, Estación Experimental de Zonas Aridas, Consejo Superior de Investigaciones Científicas CSIC; El Ardal, Murcia, (Spain), managed by the University of Murcia; Lanaja, Zaragoza, (Spain), managed by the University of Zaragoza; Var and Roussillon, (France) managed by the BRGM; Draix, (France), managed by CEMAGREF; Santa Lucia, Cagliari, Sardinia (Italy), managed by the University of Cagliari; Araba, Veneto, (Italy), managed by the University of Padua; Spata, Athens, (Greece), managed by the Agricultural University of Athens.
Some of the research sites are integrated in networks, as in the case of the Spanish RESEL, (Network of measurement and monitoring of erosion and desertification) which is a cooperative project to maintain, co-ordinate, and harmonise the existing experimental fields. It is coordinated by the General Directorate for the Conservation of the Nature of the Ministry of Environment and integrates 47 experimental fields distributed around the country and managed by 11 universities and 7 CSIC (Spanish Council of Scientific Research) centres.
Experience has shown several problems to appear relatively frequently during the implementation and operational phases of work. Some of these are of a general nature and some others are specific technical problems that have arisen in connection with the experimental design, the operation of instruments and data storage, management and interpretation. Such technical problems are detailed in the Chapters 2, 3 and 4 of this document.
Among the general problems that may arise during the implementation and operation of the experimental areas several must be emphasised.
One of the main concerns in establishing an experimental area is to secure the control of the land throughout the duration of the experiment, either through ownership or by legal agreement. This is a critical factor in a middle- to long term activity such as measuring soil erosion. Stable land ownership, whether public or private, is essential. Public ownership does not always guarantee stable conditions, particularly if two or more contending agencies of government claim jurisdiction; often contention arises because of poor goals, weak justification or inadequate support.
Another aspect to consider is related to the person responsible for the experiment. Too often the researcher who plans and begins an experiment will not remain to finish it, therefore, he must leave a clear record of methods and procedures. Also it is important to appoint a unique specific agency and person in charge of the experimental area. The design and operation of experiments through interdepartmental committees, with no agency or project leader clearly in charge, has invariably led to failed objectives because of confusion about responsibilities and credits arising out of the inevitable cross-purposes between agencies. (Hewlett, 1982).
Another major concern is the size of the experimental watershed. Once more this is connected to the research objectives and the management questions which have to be answered.
The integral monitoring of the erosion cycle requires an area large enough to contain at least the first order perennial stream. Something that might require tens of thousands of hectares in some semi-arid Mediterranean environments, or simply be impossible to achieve. As a compromise solution the watershed should contain at least a clear and defined ephemeral stream where the channel phase of the erosion cycle can be monitored.
The question of the watershed size is also connected with probably the most difficult problem related to erosion measurement in arid and semiarid lands, i.e. the irregularity of the rainfall regime, and the fact that erosion phenomena increase with the intensity of the rainfall. Large storms with a long recurrence interval, although difficult to measure, are critical in explaining the erosion cycle. The solution to such dilemmas should be sought in the context of each particular experimental initiative. It is frequently necessary to identify the specific classes of discharges that are meaningful to the scientific response that is to be provided.
The size of the plots is obviously also a consideration. Again this should be adapted to the experimental objectives. When possible, it is advisable to use existing standard dimensions. For this purpose the USLE plot provides a "universal" reference.
Data management storage and exploitation could turn out to be a nightmare to the researcher, particularly if he is not the person who initiated the project and the person formerly responsible left an inadequate documentation and record of the work done.
A clear, reliable, updated, accessible and thoroughly documented system of data storage and management is vital for the long-term success of the research.
Finally, a pressing technical problem is the measurement of bedload. Bedload measurement is a matter of great uncertainty in Mediterranean "rambles", in mountain torrents and in big rivers. If bedload measurement is one of the priorities of the experiment, it exerts a strong control on the whole experimental design; the more so the larger the size of the tributary basin and channel.
From several designs proven in practice, the stilling pond (Hewlett, 1982) and the Araba torrent station (Fattorelli et al, 1986) provide the most effective solutions for total load measurement.
1.2 Justification for implementing measurements of erosion processes
Soil is the basic resource interfacing most of the ecological and socio-economical processes. This simple statement justifies in itself prioritising the soil in terms of the attention it is given with respect to natural resources conservation and economic development planning.
Soil formation takes centuries, millennia or even longer periods, while soil destruction occurs in years, months, days and even seconds. An instant compared with the time necessary to reclaim it, if this is indeed possible.
The United Nations 1990 Global Assessment of Soil Degradation (GLASOD) study presents the extent of soil degradation in the world. Nearly one sixth of the world's vegetated area has suffered some degree of degradation during the last 50 years. More than three quarters of this degradation is caused by agriculture and livestock production or by converting forest to cropland.
The degradation processes are diverse: salinisation and water logging on poorly managed irrigated lands; compaction caused by the use of heavy machinery; and pollution from the excessive application of pesticides or manure. But erosion is by far the most common type of land degradation accounting for 84 percent of affected areas, according to the cited study.
In this context, the need for clear and accurate soil erosion measurements is critical as a reliable basis for the development of prevention and reclamation plans, particularly considering the fact that soil erosion is a silent and complex phenomenon arising from the combination of multiple factors.
Such complexity is at the roots of the limitation of mathematical models, making them of relatively limited application away from the particular conditions existing where they were established.
Soil erosion measurement provide a similar reliable basis for water resources conservation and planning. Both water quality and the water regime are greatly affected by erosion. The fine soil particles that travel farther in suspension are the most extensive pollutant and serve as a vehicle for other chemicals and organic contaminants. Soil erosion is intimately related to floods, particularly in the Mediterranean, to the extent that any study of flood genesis and prevention in the region should consider the erosion phenomena and requires reliable field data on them.
It has been mentioned that efficient erosion control depends on the knowledge of soil erosion rates and mechanisms. Such needs are particularly felt when dealing with complex phenomena, frequently subject to trivial "solutions" that can cause unforeseen effects. In this sense the erosion measurement is a valuable tool for the design of erosion control measures, through the accurate evaluation of the efficiency of management methods applied in soil protection and reclamation.
In spite of the relatively large attention that has been given to soil erosion measurement, an important gap that can be thoroughly identified is the lack of homologation and standardization of these measurements. It is critical that these be made comparable and completely useful both for national and international purposes.
One of the basic objectives of this project is to establish a feasible, real world and real-scale contribution to the solution of the lack of standardized and coordinated international soil erosion measurements. This must be meaningful in the context of a global environment affected by global processes which require a global, coordinated standardized monitoring as the only way to formulate co-operative and coordinated response that present circumstances demand.
Finally, the basic prerequisites for the organization of erosion measurement activities are present in all of the Mediterranean countries. From the institutional framework to the technical capacity, all the necessary elements are easily identifiable in the region.
2.1 The erosion-sedimentation sequence
2.2 Rainfall-induced erosion phenomena
2.3 Measurement sites
2.4 Devices and instruments for measuring sediments
2.1 The erosion-sedimentation sequence
The generic term "erosion" includes a series of natural events, each one being the consequence of the previous one, and affecting all of the parameters to be considered in watershed management. These events, which form a sequence in time, are the following:
• causative factors of erosion: chemical action, temperature variations, frost, topography, plant cover and human influence such as: road or railway constructions, industrial projects, mining, sewage wastes, waste dumps, misuse of soils, water and agricultural land. All of these factors, alone or combined, provoke the weathering and the degradation of the surface layers, which are progressively eroded by the impact of rainfall, removed and transported away;
• erosion which reveals itself in many different forms: gullying, mass movement of soil or landslides, flood erosion, sheet erosion, stream channel erosion, etc.;
• the sediment transportation from the eroded site by a vector such as water, wind, snow, or glacier, etc., to drainage channels and downstream by channel flow;
• the sediment deposition, sedimentation or siltation on land, in a water course, lake, or sea, etc. In turn, this last stage can again be the starting point for a new erosion sequence due, for instance, to a change of climatic conditions and/or tectonic events which may reactivate erosion processes remobilizing the formally deposited sediments.
Each of the above events characterising the erosion sequence is a phenomenon which develops in a space having from zero to three dimensions: 0-D point (e.g. location of a sediment deposit), 1-D line (e.g. sediment transport), 2-D surface (e.g. sheet erosion) and 3-D volume (landslide). Each such event is the consequence of previous ones. It is, therefore, understandable why the measurement of "erosion" is so complicated, difficult and rather unreliable. In contrast with other main hydrological variables, such as rainfall, streamflow, snow, etc., the erosion sequence is a one-way process in the human time scale and thus cannot produce two similar events since sediment material sources, once eroded are not renewable. Consequently, the erosion sequence is basically not stable and generates the cause for its transformation, since erosion, in the long run, will transform the site morphology and topography and consequently, its erodibility. This is precisely what takes place very quickly when human pressure on the land results in massive removal of the protective vegetation cover: erosion increases drastically, as well as the sediment transport in river courses; eroded materials build up large sediment deposits downstream; finally erosion declines when bedrock becomes directly exposed and through the removal of all of the erodible surface material and the area becomes sterile for agricultural purposes.
With such a chain of complex and variable phenomena, both in time and in space, one might question how to measure, what and when? The following sections attempt to give an accurate answer to measuring the erosion process.
2.2 Rainfall-induced erosion phenomena
The erosion-sedimentation sequence is complex due to the great number of physical elements, function of both time and space, that conditions each step of the process. In Mediterranean areas, the complexity is emphasized due to the generally greater variability of these elements than in more temperate areas.
2.2.1 Rate of erosion
Extreme runoff events are of fundamental importance in the erosion-sedimentation processes. For example, a drainage channel may serve as a major sediment storage system during a rather long period when moderate runoff peaks cause a progressive rise in the longitudinal profile of the drainage channel system. The incoming sediments are delivered at such a rate that the 'normal' flows are incapable of coping at a sufficient rate, for a variable period of up to several years. When an exceptional flood occurs, a large proportion of the accumulated sediment may be removed at once, and the bed may even be scoured below its previous lowest level. This demonstrates that a correlation between erosion rates and sediment transport in a river downstream cannot be significant over short periods of time, and that there is certainly a varying time lag between the two events. It is likely that the relationships established are much more complicated than simple proportions.
Quantitative determination of the rate of erosion within a watershed has been the subject of intensive research and data collection over many years, but mainly on small trial test plots, often with rainfall simulation. A number of equations and methods of computation have been developed, amongst them the "Universal Soil Loss Equation" (USLE) which computes empirically the quantity of material eroded, usually expressed in weight per unit area per year, as a function of a series of factors including:
• Climate: rainfall intensity, duration and frequency; temperature variability both seasonal and daily; and cycles of frost, thaw, wetting, and drying; etc.
• Hydrology: type and intensity of surface water runoff, assessment of water carrying capacity of sediments at all stages, from rain drops impact (splash) to the rivers and tributaries of the main catchment drainage pattern.
• Geology: rock formations, grade of weathering, hydrodynamic properties;
• Topography: slopes, relief, altitude, etc.
• Plant cover: nature and density of species; seasonal variability
• Human factors: agriculture, urbanisation, engineering works, deforestation.
Analytical studies have so far not reached the point where satisfactory mathematical relationships between rates of erosion and their causal factors can be derived, but some significant trends are, however, apparent in some regions of the world.
Under a basically similar climate, topography, hydrography and soils, small drainage areas with their tendency toward a wide diversity of land use have a greater range in rates of sediment production than large drainage areas. As the drainage area becomes larger, pronounced land-use differences and, consequently, great variations in erosion tend to be smoothed out. Therefore, the range of production rates tends to become smaller. Generally, also, the larger the drainage area the higher will be the percentage of sediment permanently deposited and thus the smaller will be the mean rate of sediment production per unit area, as measured at the drainage outlet of the watershed.
2.2.2 Sediment delivery ratio
The sediment delivery ratio is expressed over a long period of time and is defined as the percentage between the sediment transported by a river and the total quantity of erosion material in movement, both relative to the drainage area at one particular section. Measurements undertaken in existing reservoirs show that the sediment delivery ratio decreases with increasing watershed areas. As an indication, the sediment delivery ratio ranges from 20% to 90% in very small watersheds of less than, say, 2 km2 and 3% to 15% in watersheds comprised between 100 and 1000 km2.
Sediment not transported by the stream from the watershed area to the sediment measuring section is deposited in channels, on flood plains, terraces and alluvial deposits. The erosion process and the delivery ratio cannot, at present, be quantified with sufficient precision for project design purposes.
2.2.3 Trap efficiency
The concept of sediment delivery ratio should not be confused with that of "trap efficiency" which is the percentage between the sediments trapped in a reservoir and the total sediments that enter the reservoir. The specification of this ratio is required in order to estimate the rate of siltation of a reservoir by the water and sediment discharges of the drainage system at the site. Trap efficiency depends on:
• the sediment size-distribution, which determines the fall velocity of suspended material and which increases with the size of the particles;
• the time used by an elemental volume of inflow water with suspended sediments to travel from the inlet to the outlet of the reservoir;
• the temporal variability of flow into the reservoir.
Several authors have proposed empirical relationships based on data derived from surveys of existing reservoirs. However, these relationships are rather site-specific as they do not take quantitatively into account the particle size distribution.
2.2.4 Carrying capacity
The erosion-sedimentation sequence is characterised by:
• hillside/slope erosion processes which provide detrical sediments and colluvial debris to the main collector of the drainage pattern;
• the hydrodynamic processes which evacuate most of the debris depending on the sediment carrying capacity of the drainage network.
Specific characteristics of transported material (specific weight, grain size distribution, shape of particles, etc.) and those of the river bed (roughness, slope, cross-sectional profile, etc.) determine the transport carrying capacity of a river. Transport carrying capacity varies along the course of a river as the velocity distribution changes with cross-sectional variations. It varies also with time at a given cross-section according to the water discharge. When the sediment load is less than the carrying capacity of the river at a given section, the water flow has some energy surplus which is utilised to scour the river bed, provided the characteristics of the river bed material satisfy the transport conditions. On the contrary, if, at a given section, the carrying capacity is less than the actual sediment load, a part of the material is deposited. In general, a river has a trend to scour under high discharge and to deposit under low discharge conditions: this is explained by the variation of the carrying capacity with discharge, and is evident with high turbidity during floods and clean water at low flow.
A river or water course is said to be in hydromorphological equilibrium when the uptake of sediment from the bed is approximately compensated by sediment deposition. On the contrary, a river is unstable when it follows a trend towards equilibrium either by eroding or depositing channel sediments. For example, the Po river in northern Italy has become unstable due to the accretion of its bed in the flat valley, upstream of its confluence into the Adriatic sea. This phenomenon appeared after the bed had been corrected and dykes had been constructed on both banks for flood control purposes. In some stretches of its low course, the river bed bottom has risen with sediment deposits even higher than the surrounding flood plain.
When considering the overall erosion/ transport/sedimentation sequence, two main representative measurement sites have to be identified:
• Hillside and interfluvee areas, where land erosion processes generate detrital material and debris; and
• Main water flows networks and drainage patterns, where detrital from the slopes is in transit and then deposited by the fluvio-alluvial processes and dynamics.
Many instruments and techniques have been developed for field measurement of the processes of erosion/sedimentation sequences, basically soil erosion, sediment transport and sediment deposition:
• measurements of sediment removed from experimental plots or small watersheds, installed or part of hillside, by surface runoff;
• measurements of sediment transport in water courses, either suspended or moving by dragging and saltation on the streambed;
• measurements of the volume and density of sediment deposits.
Sites suitable to measure the above phenomena are discussed below.
2.3.1 Erosion test plots and small watersheds
Small plots, a few hundred square meters, or small natural watersheds, a few hectares in size, are used to study erosion rates of various soil types which represent a land area of a specific topography and vegetative cover. Plot shape can range from a small rectangular area to a naturally shaped watershed. Plot length should represent the length of slope on which soil losses occur. The length commonly ranges from 10 to 100 m and the width may vary from 2 to 10 m. Heavy metal sheets may be used to delineate small plots. They are easily removed when cultivation of the plots is required. Earthen ridges may be used on plots wide enough to operate equipment normally used in farming operations.
Small natural watersheds usually do not require borders except dikes at the lower edge to direct the flow to the point of measurement.
Plots of all sizes are equipped with a trough or other device to collect the soil-water mixture in the sampling device. Sheet metal is used for the collecting trough on small plots. A concrete channel or earth dike may be used on large plots and small watersheds.
2.3.2 Sediment sampling stations in natural rivers
Sites used as a hydrological station for measuring water discharge measurements in natural rivers are adequate for suspended sediment measurements. In addition to the usual requirements of access and availability, a straight channel reach with rather uniform sediment distribution and velocity is desirable. Consideration is also given to the construction of cableways and bridges and to the installation of the equipment for sampling suspended sediment concentration. As yet, equipment is not available to sample the entire depth of flow in a channel. Depth integrating samplers can only sample the flow from the water surface to about 10 cm above the streambed. This can be an important part of the total water depth under low discharge. Therefore, artificially or naturally turbulent sections which would spread uniformly suspended sediments can provide an approach to total load sampling. Also, for rather small streams, many flow-measuring control structures such as weirs, notches and flumes can be designed to provide some overfall in order to sample the total sediment load of the flow directly into a rather large-size container. Road box culverts and other hydraulic structures may also be used to sample the sediment concentration of a stream discharge.
2.3.3 Artificial reservoirs
Existing artificial reservoirs are often used for sedimentation investigations as they usually act as an "integrator" of the erosion status of the typical land resource area which constitutes the watershed. Studies are generally directed towards determining the quantity, characteristics and distribution of sediment deposits measured periodically by means of topographic surveys of the reservoir bottom. Conventional measurements of sediment inflow and outflow are also required in order to determine the sediment trap efficiency of the reservoir. Essential factors are reservoir size, shape, capacity, inflow and outflow rates and volumes and watershed characteristics.
In order to survey the amount of sediments deposited in the reservoir during a given period (e.g. one year), it should be emptied if conventional topographical instruments are to be used (level, theodolite). Otherwise, a boat equipped with a sonar or with a graduated weighted tape is used provided that the exact position of the boat is determined accurately, either from the boat or from the shore. After survey, the actual topography of the bottom of the reservoir is compared with that of pre-construction in order to determine by difference the total volume accumulated since the date of construction of the reservoir.